results that he did not anticipate, and which he considered to be of sufficient importance to justify him in presenting them to the Society. The object originally aimed at in its construction was to obtain an exact measure, by weight, of the actual amount of deflective force which the current exerts upon the magnetic needle. The instrument constructed for this purpose is represented in Fig. 1. The coil consists of a total length of 1,660 feet of No. 22 copper wire, weighing rather more than 6 lbs., and divided into four parts, the ends of which are brought out and connected with their respective terminals G G, so that they can be used separately or as one coil. The needle with which the first experiments were made consisted of a small rectangular steel bar, 14 inch in length, rather less than th of an inch in breadth, and about half the thickness of a shilling. It weighed exactly 18 grains, and, when magnetized, its lifting power was 44 grains, or nearly 2 times its own weight. The index M, which is 9 inches in length, weighs only 2 grains. A small brass pulley, th of an inch in diameter, is fixed upon the axis between the index and the supporting screw L. The balanced lever EH consists of a thin slip of hard spring-brass, placed edgewise for strength, and tapered, for lightness, towards the end of the long arm A H. The short arm A E is loaded to act as a counterpoise, and to this arm a scale-pan C is suspended, at a distance from the fulcrum, equal to exactly th of the length of the other arm. It carries also at its extremity a thin horizontal projection, which vibrates between two screw-points D, D', and by which, with the aid of the wooden foot-screws of the instrument, the lever can be always exactly levelled when balanced. The fulcrum B is supported on a stout brass bar F, which is firmly held in its place by means of the screw I, and can be removed at pleasure. When it is desired that the needle shall have liberty to move in both directions, the extremity H of the long arm of the lever is connected with the needle by a slender wire suspended from a very fine thread, fixed to the upper part of the pulley, and carried down on both sides of it, as shown in Fig. 2. The arm A H is divided into 10 equal parts, each of which is subdivided into tenths; and, estimating the poles of the needle to be at a distance of about 4th of its total length from the extremities, the diameter of the pulley is so adjusted that a weight of 100 grains, suspended at the distance of one of the large divisions from the

Fig. 2.

fulcrum, acts with a force of 1 grain at the poles of the needle; suspended at division 2, it acts with a force of two grains; at 2.5, with a force of 24 grains, and so on.

The following table exhibits the results of the first series of experiments made with a small Grove's battery, the platinum plates of which expose only two inches of surface, and having the zinc plates immersed in a saturated solution of chloride of sodium. It is a striking characteristic of Grove's battery that it slightly increases in force after being some time in action, and it would have been preferable, therefore, to use a Daniell's, on account of its remarkable constancy, but the writer had not a sufficient number in series. The fourth column indicates the weight required to bring the index of the balance galvanometer back to zero; the fifth column expresses the same weight reduced to the force which it exerts at the poles of the needle, but increased in each case by half a grain, to compensate for the small preponderance given to the long arm of the lever in order to keep the needle vertical when not deflected by the current :

It will be seen that up to six pairs, the numbers in the fifth column, expressing the force of the current in grains at the poles of the needle, vary very nearly in the same ratio as the tangents of the angles of deflection on the tangent galvanometer reduced to a comparable form in the sixth column. With twelve pairs, however, the weight required to balance the current is 495, or very nearly fifty grains; whereas, according to the tangent galvanometer, it should not have exceeded forty grains. Reflecting on this anomaly, the writer could only arrive at the conclusion that the needle, surrounded by a very powerful current in such a large coil, ceased to act as a permanent magnet, and was temporarily charged with a higher magnetism induced by the current itself. Subsequent experiments completely confirmed this conclusion; and he was led to examine the subject more minutely by observing that the needle, after being several times subjected to the

action of the current from twelve pairs, appeared to have lost its permanent magnetism; for he afterwards found repeatedly that when the index was brought back by successive increments of weight to 40°, the smallest possible additional weight sent it back to zero. Before taking out the needle to ascertain this, he submitted it a second time to the action of six and twelve pairs, with the following results :

Number of Pairs.

6

12

Current Force by Tang. Galvan.
22.64
39.64

Weights supported.

7.5 grains.

24.5 ""

Whereas it will be seen, by referring to the preceding table, that originally it supported 22·50 and 49-50 grains respectively. The needle was then taken out, and was found to have almost entirely lost its magnetism. It had originally lifted forty-four grains; it was now only with the greatest precaution that it lifted two grains. But 7.5 : 24.5 = 1 : 3.27 (39.64)21: 3:06.

and (22-64)2

The writer had therefore little doubt that if not merely demagnetized, but formed of soft iron, the needle, when placed in a favourable position, would turn with a force proportional to the square of the current; whereas it plainly appears from Table I., that so long as its permanent magnetism is sufficient to resist the inducing action of the current, the needle is deflected with a force simply proportional to the current.

To determine this interesting question, two new needles were constructed, similar in shape and size to the former, but somewhat lighter, each weighing only 17.25 grains. The one was of steel, tempered to the hardness of glass; and was magnetized till it lifted with some difficulty forty-three grains; the other was of soft hoop-iron, well annealed. With these needles the following results were obtained from experiments conducted very carefully, and using, for greater accuracy, a single-thread suspension :—

:=

A glance at the above table will show that, with the permanently magnetized needle, the numbers in column 4, expressing the deflective force of the current in grains, are very nearly proportional to the numbers in column 5, expressing the simple ratio of the tangents or quantities of current; whereas with the soft iron needle they are nearly proportional to the squares of the same quantities, reduced to a comparable form in column six. In both cases the only marked deviation coincides with the powerful current from 12 elements of the battery, in which case the steel needle, evidently acting under the superadded influence of induced temporary magnetism, is deflected with a force which exceeds the estimated amount by 47 grains, whereas the soft iron needle, under the same current force, falls short of the calculated amount by 114 grains. The last effect is probably attributable to the fact that, as the needle approaches saturation, the law of the squares gradually merges into the law of simple proportion, The writer regrets that he had not at command sufficient battery power to put this point to the test of decisive experiments, but hopes to do so shortly with a Daniell battery of 50 or 100 elements. In the meantime the results above given, having been arrived at with great care, and amply confirmed by experiments several times repeated, appear to establish very conclusively the following principles :

:

1. A permanently magnetized steel needle, suspended in the middle of a galvanometer coil, is deflected with a force simply proportional to the quantity of current transmitted, so long as the current force which acts upon it is not sufficient to impart temporarily a higher magnetism than that which it permanently possesses. Beyond this point, the deflective force exerted on the magnetized steel needle increases in a somewhat higher ratio than the current, and therefore the accuracy of any form of galvanometer can be trusted only within certain limits of current force and of length and proximity of coil.

2. A pure soft iron needle, suspended at an angle of about 40° to the direction of the current (the angle varying according to the shape of the needle), is deflected with a force which, within certain limits of current power, is very exactly proportional to the squares of the quantities of current. Beyond these limits the deflective force exerted on the needle increases in some constantly diminishing ratio lower than that of the squares of the current.

3. The action of the current in deflecting a magnetic needle is precisely the same action, and follows the same law, as that which it exerts in magnetizing a bar of soft iron. The amount of magnetism actually imparted to a bar or needle of soft iron is directly proportional to the quantity of current; for the force with which a soft iron needle is deflected under different currents is not proportional to its tempo

rary magnetism in each case, but to the product of its magnetism multiplied by the force of the current. By increasing the force of the current, two effects are produced; in the first place, the magnetism of the needle is increased in the same proportion; and secondly, the increased current acting upon this increased magnetism deflects the needle with a force proportional to the product of the two, or in other words, proportional to the square of the actual quantity of

current.

It only remains to add the results of two series of experiments, shows ing the very striking difference between the defective forces exerted upon the two needles at different angles of inclination. Table III. shows the increasing weights required to balance the needles at angles successively diminished by 10°; Table IV. exhibits the effect produced by successive additions of weights, equivalent to a force of ten grains acting at the poles of the needles. In both cases the battery power employed was twelve small Groves, but the current declined, in the course of the experiments, from 47° to 45° on the tangent galvanometer, which accounts for the fact that the maximum weights supported are less than in the earlier experiments recorded in Table II. In working out Table III., the weight employed (1,000 grains) was simply advanced along the lever, and its reduced amount at the poles of the needle noted, when the index, in gradually retreating, pointed to the successive angles specified. The results in Table IV. were obtained by moving the weight from one to another of the successive divisions, marked 1, 2, 3, &c., on the lever; and the differences of the angles vary, as might be expected, in nearly the reverse order of the differences of weights in Table III :—